Prevalence of Dyspepsia in Iran: A Systematic Review and Meta-analysis

1Department of General Surgery, Faculty of Medicine, Ilam University of Medical Sciences, Ilam, Iran 2Professor of Medical Bacteriology, Molecular Biology Research Center, Systems Biology and Poisonings Institute, Baqiyatallah University of Medical Sciences, Tehran, Iran 3Department of Psychiatry, Faculty of Medicine, Ilam University of Medical Sciences, Ilam, Iran 4Faculty of Medicine, Qazvin University of Medical Sciences, Ilam, Iran 5Faculty of Medicine, Ilam University of Medical Sciences, Ilam, Iran


Introduction
Dyspepsia is a very prevalent gastrointestinal disease and its prevalence has been reported to be 40% in populationbased studies. 1 The Rome II Criteria for the diagnosis of functional dyspepsia (FD) include: FD for more than twelve weeks, which does not need to be continuous, within the past twelve months with: (a) Constant or repeated symptoms (any kind of discomfort or pain focused in the upper part of the abdomen); (b) No sign of organic illness (on endoscopic examination) that is probable to describe the manifestation; and (c) No sign to demonstrate that FD is just improved by defecation or related to the beginning of a change in form or frequency of stool (i.e., not irritable bowel syndrome [IBS]). FD includes (1) ulcer-like FD: Discomfort or pain concentrated in the epigastric part of the abdomen is the main manifestation (the most annoying); (2) Dysmotility-like FD: A disagreeable or irritating painless sensation concentrated in the epigastric part of the abdomen is the main manifestation; this sensation is perhaps identified by or associated with early satiety, fullness of upper abdominal, nausea or bloating; (3) Unspecified FD: patients whose manifestation do not meet the prerequisites for dysmotility or ulcer-like FD. 2 According to the Rome IV criteria, which were recently revised, 3 FD is characterized by: (a) Constant or repeated dyspepsia for above three months within the past six months; (b) No sign of a probable organic reason of the endoscopic examination; and (c) No evidence that the dyspepsia is only ameliorated by defecation or any relation with stool abnormalities. This last criterion serves to omit IBS as a possible reason of the manifestation, although about 30% of patients with FD suffer from IBS.
Although clinical assessment is needed to specify if dyspepsia is FD or structural in a person, endoscopic examination shows that most are FD. FD is a complex issue with probable factors such as visceral hypersensitivity, immune activation, brain-gut axis dysfunction, and delayed gastric emptying and Helicobacter pylori infection. [4][5][6][7][8] Although FD is a serious issue worldwide, most of the reports come from Western countries. Dyspepsia is usually of structural type in developing countries, while in developed countries, FD is more prevalent. 9,10 Although FD is not a life-threatening illness which usually does not require surgery and has not been shown to decrease survival, 11 it disrupts the quality of life significantly and is associated with high rates of absence from work, poor efficiency, fewer activities around the house, and higher medical expenses, 12,13 demonstrating that dyspepsia in general causes many mental health complications for the population. 14 Many researchers have investigated the prevalence of dyspepsia in the Iranian population, but the results are not consistent. [15][16][17][18][19][20][21][22][23][24][25][26][27][28] Thus, a coherent survey of all documents and combining them can provide a more comprehensive picture of the dimensions of this illness in Iranian community. The results of this meta-analysis can provide a more accurate estimate regarding the effect of therapies or risk factors for the illness or other results through combining different studies. [29][30][31] The present research was carried out to assess the prevalence of dyspepsia in Iran.

Study Protocol
This research was carried out according to Metaanalyses Of Observational Studies in Epidemiology (MOOSE) guideline 31 and reporting was performed according to the Preferred Reporting Items for Systematic Reviews and Meta-analysis (PRISMA) protocol. 32 This meta-analysis was recorded in PROSPERO with code (CRD42019148610).
Two authors performed each step of the study independently. A third author resolved the inconsistencies in each step of the research and they were confirmed through discussion.

Search Strategy and Selection of Studies
A systematic search was conducted in July 2019 using international databases, including PubMed/Medline, Science Direct, Web of Science (ISI), Cochrane Library, Scopus, EMBASE, EBSCO, CINAHL, and national databases such as Magiran, Iranian Research Institute for Information Science and Technology (IranDoc), Scientific Information Database (SID), Barakat Knowledge Network System, Regional Information Center for Science and Technology (RICST), and the Iranian National Library, as well as Google Scholar.
The key terms are introduced in Table 1. We selected the "All Fields" option as the field of search. We combined the terms "descriptive", "population sentences" and "outcome" using the Boolean operator "AND".
To identify all potential studies, the reference lists of all found articles were manually reviewed.

Inclusion and Exclusion Criteria
The inclusion criteria were all cross-sectional articles on the prevalence of dyspepsia in Iran in both Persian and English. The exclusion criteria were: (1) non-adult samples (more than 10% below the age of 15); (2) non-random samples; (3) non-Iranian subjects; (4) non-related subject; (5) sample size below 100 participants; (6) participants not from the general population or community; (7) case reports, case series, congresses, letters to the editor without quantitative data, reviews, and dissertations.

Study Selection
A reviewer (M.A) reviewed the titles and abstract of all detected reports and the full text was then assessed. The same action was repeated by a second researcher (M.K) independently. Eventually, any disagreement was discussed and resolved in the presence of a third researcher (M.S).

Data Extraction
The following data were recorded on a spreadsheet from each study: first author's name, publication date, year of study, mean age and standard deviation, how the data were collected, the diagnostic technique for dyspepsia, duration of symptoms for dyspepsia and prevalence of dyspepsia. At the end, the obtained data were imported into Excel.
For overlapping articles, we emailed the corresponding authors to determine the original article, and in case we did not receive any answer, we chose the article with the highest number of subjects for cases that overlapped.

Quality Assessment
Since all articles included the cross-sectional studies, the quality of the each study was assessed using the modified form of Newcastle Ottawa Scale. 33 The study quality was categorized as follows: poor quality = 0-5; medium quality = 6-8; high quality = 9-10. Articles with poor quality were excluded.

Statistical Synthesis
All analyses were performed using the Comprehensive Meta-Analysis software version 2. Heterogeneity was evaluated using I 2 and chi-square tests. Heterogeneity was classified as follows: I 2 values 0%-25%, 26%-50%, 50%-75% and greater than 75% showed low, medium, high and very high heterogeneity, respectively, with P value <0.10 considered statistically significant. 34,35 In cases of high levels of heterogeneity among studies, we used the random effects model. The male/female odds ratio (OR) was used to show the impact of gender on dyspepsia. For this goal, the prevalence of dyspepsia in males, females and the total sample size was evaluated. Subgroup analysis was performed according to areas, population groups, study period, and diagnostic technique. Mixed effects metaregression was utilized to assess the relationship between study time and the prevalence of dyspepsia. Subgroup differences were also tested. 36 Sensitivity analysis also was done. Publication bias was evaluated by an in-depth analysis of the funnel plot, using the proposed correlation tests of Egger and Begg's. 37,38 P value less than 0.05 was considered statistically significant.

Search Results and Features
Overall, 1460 articles were retrieved. Of these, 1446 studies were excluded because of: duplicate publications (n = 730), unrelated literature (n = 651), non-epidemiological studies (n = 17), non-random samples or studies on children (n = 28), participants not from general population or community (n = 18), as well as case reports, case series, congresses, letters to the editor without quantitative data, reviews, and dissertations (n = 2). Finally, fourteen studies remained for the meta-analysis ( Figure 1). The mean age of samples was 37.41 years (95% CI: 33.70-41.11). Three, two, one, two, four, one, and one studies were conducted in Tehran, Fars, East Azarbaijan, Hamadan, Isfahan, Kurdistan, and Kerman provinces, respectively. Ten, two, one, and one studies were conducted on the general population, healthcare workers, outpatient clinics, and blood donors, respectively. Four and ten studies had low and medium quality, respectively. Table 2 demonstrates the characteristics and general information of each study.

Prevalence of Dyspepsia
The prevalence of dyspepsia by combining fourteen studies with a sample size of 54 118 subjects was 14.6% (95% CI: 9.6-21.7) (Figure 2A). The minimum and maximum    Figure 2B).

Prevalence of all Types of Dyspepsia
The prevalence rates of ulcer-like dyspepsia in five studies with a sample size of 8,000 subjects, dysmotility-like dyspepsia in five studies with a sample size of 8000, and unspecified dyspepsia in four studies with a sample size of 4,483 were estimated to be 12.1% (95% CI: 5.2-25.7), 9.7% (95% CI: 4.9-18.4) and 17.0% (95% CI: 7.8-33.4), respectively ( Figure 3).

Sensitivity Analysis
Sensitivity analysis showed that the total estimates are robust for prevalence of dyspepsia and for males and females ( Figure S2, Supplementary file 1).

Discussion
This research is the first meta-analysis addressing the prevalence of dyspepsia in Iranian people. The prevalence of dyspepsia was estimated to be 14.6% based on 14 studies. In a review study, the global prevalence of dyspepsia was found to be 31%, while indicating different prevalence in various areas such as North America (30%), South America (29%), North Europe (29%), South Europe (37%), South Asia (49%), Southeast Asia (27%), and Australia (25%). 39 Various meta-analyses have been conducted on the prevalence of dyspepsia in special disorders, and the prevalence of dyspepsia was 43.9% in GERD patients, 40 65.7% in patients with cholecystolithiasis 41 and 9.55% after acute gastroenteritis. 42 In another review, the global unexamined dyspepsia prevalence was 29%. However, in studies in the Middle East (including Iran), it was reported to be 15%. 43 A systematic review and meta-analysis study in Iranian population the daily, weekly, monthly, and overall prevalence of GERD symptoms was 5.64%, 12.50%, 18.62% and 43.07%, respectively. 44 The present meta-analysis has many strengths. We utilized a universal and repeated search to increase the ability to detect all related papers. All phases of the study were independently carried out by two authors, and the disputes were resolved by a third author. We made contact with the first or corresponding authors to make sure we omit duplicate publications, and to gain additional data in certain cases. A random effects model was used to merge the data to prepare dependable estimates of dyspepsia prevalence. We also performed subgroup analysis to identify the reason for heterogeneity and to assess publication bias. Finally, the limitations of the present study are associated with the search in local databases, which lack the potential for combined search. We also excluded studies that were performed on special patients or had non-random sample size, so that the estimate would be attributable to the general population. Moreover, the limitations of the initial studies were about data collection, as some of them used questionnaires and some used interviews or both. Of course, more personal methods, such as interviews, which are the dominant methods in studies included in the meta-analysis, may overestimate the prevalence, but more impersonal techniques, like completing a questionnaire form on a website, might be more accurate. The absence of articles that show the prevalence of dyspepsia for some geographical regions such as eastern and northern Iran may be considered as other limitations. Moreover, there was significant heterogeneity across all studies in our analysis, and considering the accessible data, we could attribute this difference to the geographic area (P < 0.001), population groups (P < 0.001), and duration of symptoms (P < 0.001) by subgroup analysis. Nevertheless, it seems that due to other cultural differences between the population groups, such as diversities in lifestyle and ethnicity (given that Iranian people have comprise several ethnicities with different cultures), 43 the accuracy and robustness of data collection could not be evaluated based on the available data.
The combination of epidemiological studies in Iran in this research showed that the prevalence of dyspepsia was significantly higher in women compared to men (OR = 1.52). In a meta-analysis at the global level, the unexamined dyspepsia prevalence was significantly higher in women compared to men in the Middle East (OR = 1.91). However, it was not significant in African countries, Southeast Asia, South America, Central America and Australia. 45 In the present study, the lowest incidence of dyspepsia based on the diagnostic technique was observed for the ROME III criteria (9%). These criteria were published 12 years ago for defining dyspepsia 46 ; few studies have evaluated these criteria and no significant difference was observed between ROME III and previous criteria for dyspepsia in one study. 47 The prevalence of dyspepsia according to the study year showed an increasing trend in the meta-regression model, although it was not significant. This increase might be due to the fact that dyspepsia has no impact on mortality. 11,48 Evaluating publication bias in the results of this study showed that bias had no impact on these studies. The bias is often created in these articles, since such articles are more probable to be printed because of their positive results. However, no such bias was found about dyspepsia in this study. 49 In conclusion, the prevalence of dyspepsia in Iran is relatively high (14.6%) but is lower than worldwide estimates.